Cross-Lingual Speaker Diarization for Romanian Without Fine-Tuning: End-to-End MSDD Versus a Traditional Segmentation–Embedding–Clustering Pipeline

Cross-Lingual Speaker Diarization for Romanian Without Fine-Tuning: End-to-End MSDD Versus a Traditional Segmentation–Embedding–Clustering Pipeline. Remus-Dan Ungureanu; Mihai Dascalu. In: Electronics, Vol. 15, Iss. 4, February 2026, DOI: 10.3390/electronics15040778.